I would say two things:

First, STCKF is potentially so much faster than STCK that it is probably well 
worth it even if you have to test a flag and branch.

Second, I am going to guess that you are safe assuming STCKF is here to stay. 
In my 55 years on this platform I have seen very, very few non-privileged 
hardware instructions go away. TBEGIN and friends was a shocking exception and 
I doubt very much that it will become the rule. TBEGIN was probably problematic 
from the get-go and should not have been introduced; not so STCKF which is 
fundamentally just a dumbed-down STCK.

Also, don't forget about the option of STCKE.

Charles

-----Original Message-----
From: IBM Mainframe Assembler List <ASSEMBLER-LIST@LISTSERV.UGA.EDU> On Behalf 
Of Mario Bezzi
Sent: Thursday, April 10, 2025 11:51 PM
To: ASSEMBLER-LIST@LISTSERV.UGA.EDU
Subject: Store-Clock-Fast Facility

PoP states "The store-clock-fast facility *may* be available on a model 
implementing z/Architecture."

The "may" word forces my performance sensitive code to go through additional 
controls before using STCKF, which sort of defeats the purpose of the 
instruction. 

STCKF was first introduced in 2005 and has been available in any CPC generation 
since then. If I could rely on it being always available I would have less code 
to run and maintain.

Is there documentation stating which facilities are available in a given CPC 
family? What others do?

PoP states the same for Transactional-Execution Facility, still IBM started 
warning long in advance about its withdrawal.

Thank you!
mario

Reply via email to